Source: (2010) University of Illinois Law Review. 2010(2): 429-487.The currently dominant view of how people process evidence and draw conclusions is linear and Bayesian. Dissatisfaction with this model, as both a descriptive and normative matter, has emerged from across disciplines- especially literature and psychology. One ambition of this Article is to synthesize these criticisms into an alternative model of how people render decisions in legal settings (the "narrative" model of judging). But I seek to do more than synthesize; my other ambition is to construct the narrative model in a way that current critics of the Bayesian models would not likely endorse at first blush. The largest departure the narrative model makes from traditional, Bayesian models of legal decision making is in demanding that judgers (judges, jurors) hear full, rich stories about the people they are judging, and treat as legally relevant the emotional sympathies that hearing them inspires. In this Article, I further modify conventional wisdom by arguing that the narrative model demands that we refuse to hear the stories of those being judged when doing so might lead us to exonerate, or even just empathize, when we ought not. (Excerpt).